Mathematical Modelling for Short-Term Urban Water Demand Forecasting Using Wavelet Support Vector Regression

Abstract

This study presents an approach for short-term urban water demand (UWD) forecasting that couple’s wavelet transform with Support Vector Regression (SVR). The wavelet transform decomposes the original urban water demand time series into multiple sub-series at various frequency levels, capturing long-term trends and short-term fluctuations. These decomposed sub-series serve as inputs for the SVR model, identifying complex nonlinear relationships among diverse factors influencing water demand. The model's performance was evaluated for lead periods of 1 and 3 days, with results demonstrating that the Wavelet-SVR model outperforms the standard SVR method in forecasting urban water demand. Performance metrics, including Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE), were used to quantify the accuracy of the predictions.